AIM staff recently attended a CLE session at the Tennessee Bar Association’s Annual Convention that covered ways for lawyers to start integrating artificial intelligence into their practices.
Before the session began, lawyers in the room were asked to share their comfort level with ChatGPT. While many lawyers are familiar with artificial intelligence, they recognize that those of us in the legal profession are slow to adapt – and somewhat leery to use - this emerging technology. Much like investing, no one wants to try something beyond their scope of expertise because the results could be devasting. A federal judge in New York City recently ordered two lawyers and their law firm to show cause why they shouldn’t be sanctioned for submitting a brief with citations to fake cases, thanks to “research” generated by ChatGPT but never fact-checked.
Professor Syd Beckman of Lincoln Memorial University in Harrogate, TN, and Wesley Clark of the Brazil Clark Firm, P.C. in Jamestown, TN, provided attendees of their session with two choices: embrace ChatGPT as it enters the legal profession or ignore it, and miss out on a billion-dollar investment being created to improve our work efficiencies.
ChatGPT is a complicated game of fill-in-the-blank that we, as lawyers, will have an ethical duty to play, based on the Rules of Professional Conduct RPC 1.1 Comment 8: “A lawyer should keep abreast of the benefits and risks associated with relevant technology.” That said, ChatGPT is a machine-learning system. It may not have the same level of understanding or judgment as a lawyer and therefore may not be able to correctly provide in-depth legal analysis. So, as we begin to use AI to enhance our legal work, we need to keep a few things in mind:
What You Should Know About & Expect from Using ChatGPT
ChatGPT is not a fact-finding source like Westlaw or Lexis, although both will start offering their own AI platforms.
You must learn how to prompt and evaluate the responses AI provides you. Be specific and articulate your query request: rather than ask ChatGPT for “a demand letter,” asking it to “write a demand letter for a plaintiff alleging personal injury damages to his neck from a rear-ended car accident” will likely result in a more appropriate product. Asking ChatGPT to “write an argument to get out of a speeding ticket where the plaintiff was going ten miles over the speed limit in a school zone on a Saturday” will result in a more specific result than asking simply for an argument to get out of a speeding ticket.
Develop the skills necessary to evaluate the outputs of AI tools, checking for errors, biases, or other issues.
The free trial ChatGPT versions available now online are not the same as the pay-to-play versions that are continuously and exponentially being invested with reliability and accuracy tools.
Hint: don’t use the free versions for work-related research.
Don’t use AI out of your depth. You still need professional judgment to fact-check anything ChatGPT provides because even the paid-for versions are not always accurate. And worse, as the New York lawyers learned the hard way, ChatGPT can fabricate case law.
AI is a time-efficient tool to help you bulk up your documents so you can edit down to what you specifically need. You can feed data into the search query (say, 2,500 raw discovery transcripts) and ask ChatGPT to analyze the data for you.
Prompt ChatGPT using language that is as clear in direction as how you would speak to a first-year associate. Providing relevant content and key details will enable the search model to better understand your need and provide results that are most useful to you.
Direct the query using the appropriate tone. If you want an argument in favor of your position, say so. If you want it to write an email response in a courteous yet assertive tone, say so.
Ask AI to first draft an outline then ask it to fill in each section. Use the output as a starting point and refine your prompt based on what the AI produced. Revise and iterate on the prompts until you get the desired output.
Do not use a free access version to run a background check on someone.
Check all citations and analysis offered by AI – if the bot steers you wrong, you can have a malpractice problem on your hands.
Keep in mind that information put into ChatGPT likely won’t remain private – do not put any private, personal, HIPAA or privileged information into a third-party AI tool. Not only could you waive privilege or violate confidentiality – your inputs could end up on the internet for all to see.
If you are an employer, consider implementing an AI workplace policy to ensure your employees know what uses of AI, if any, are acceptable in your office.
Because publicly available Gen AI tools sometimes “hallucinate” (give false output) or are biased, you can run afoul of your company’s policies, including diversity policies, if you do not double-check the outputs.
Don’t use Gen AI to make employment decisions.